Developing Predictive Models for Real Estate Pricing and Marketing Trends

Real estate is a capital-intensive and data-rich sector where accurate price prediction and effective marketing can significantly improve transaction outcomes. Traditional property valuation methods, though useful, fall short in capturing complex, non-linear market dynamics. With the increasing availability of structured and unstructured real estate data and the rise of digital marketing platforms, machine learning (ML) offers promising tools to predict prices and evaluate the effectiveness of property listings. This study applies ML models Linear Regression, Random Forest, and XGBoost to a real estate dataset enriched with location features, property characteristics, and digital marketing variables. Furthermore, it incorporates insights from a primary survey of 50 respondents to understand buyer behavior, trust in AI tools, and digital channel usage. Results reveal that models like XGBoost outperform traditional regression techniques in predicting prices, while marketing features such as emotional keywords and high-quality visuals significantly influence listing engagement. The paper concludes by recommending the integration of explainable AI, digital optimization strategies, and real-time data pipelines for smarter real estate decision-making. Keywords: Real estate pricing, machine learning, digital marketing, predictive modeling, buyer behavior, AI in property valuation.